What Algorithms Tell Us About Structural Racism in Health Care
A research study on a commercial computer program used to allocate health care resources on predicted future health care costs provided a window on the ongoing pervasive impact of structural racism in our nation’s health care system.
Moreover, the research published in Science magazine reinforced how structural racism persists throughout society, including disparities in income, housing, and other social and economic factors, and the impact that has on disparities in health.
The study focuses on an algorithm used by health systems, insurers, and practitioners to predict which patients with complex medical needs should receive extra medical care. The ostensible goal is to slash costs by suggesting those patients receive “high risk management” at less expensive primary care levels.
However, the researchers found the algorithm severely underestimates the health needs of the sickest African American patients, as the Washington Post noted, by basing its data on how much medical costs are expended on white patients compared to black patients with comparable chronic health conditions.
As one report put it, “only 18% of the patients identified by the algorithm as needing more care were black, compared to about 82% of white patients. If the algorithm were to reflect the true proportion of the sickest black and white patients, those figures should have been about 46% and 53%, respectively.”
Technology is supposedly neutral. But real life isn’t.
The research team, led by Sendhil Mullainathan at the University of Chicago, found that black patients typically generate far lower medical costs than whites for a variety of socioeconomic reasons that have nothing to do with their actual health needs, including “differential access to transportation and competing demands from jobs or childcare.”
Biases, they added, are baked into the health care system, including discriminatory treatment by some white medical professionals, such as well-documented “differential perceptions of black patients.”
As the Post’s Carolyn Johnson amplified:
“In medicine, there is a long history of black patients facing barriers to accessing care and receiving less effective health care. Studies have found black patients are less likely to receive pain treatment, potentially lifesaving lung cancer surgery, or cholesterol-lowering drugs, compared with white patients. Such disparities probably have complicated roots, including explicit racism, access problems, lack of insurance, mistrust of the medical system, cultural misunderstandings, or unconscious biases that doctors may not even know they have.”
“Machines increasingly make decisions that affect human life and big organizations — particularly in health care — are trying to leverage massive data sets” in ways that “run the risk of automating racism or other human biases,” Johnson observed.
What the articles are describing is structural racism, the legacy of slavery followed by the white supremacist overthrow of post-emancipation Reconstruction which blocked reparations for slavery — cementing in economic and social inequality in income and wealth, job opportunities, housing, education, health care, and more that persist today.
In a medical system premised on profits and ability to pay, racial disparity presents in ability to afford the high cost of premiums, deductibles, and copays that disproportionately discourages African Americans from getting needed medical care. That problem is exacerbated when corporate health care decisions are premised solely on reducing costs for care, including with medical technology as the algorithm that is the focus of the research study.
That is compounded by institutional racism, as when hospital systems close facilities or cut services in low-income communities where higher percentages of African Americans live.
Another example is resource allocation, such as the New England Journal of Medicine study that found only two of 12 Chicago hospitals that meet national guidelines of cancer care are located in the city’s predominantly black South Side. The result is black women in Chicago were almost 40 percent less likely than white women to have recommended mammograms — a reason they would be spending less on care.
National Nurses United has long warned of the serious deficiencies of health care technology, especially when algorithms dictate who gets proper care and who doesn’t. And we have seen how dangerous it can be when strict protocols ordered by medical institutions or insurance companies based on clinical software overrides the ability of RNs and physicians to use their professional judgment to diagnose and determine levels of care based on individual patient need, not large data sets.
Far more is needed as well to redress the type of structural racism exposed in this study. Medicare for All, that guarantees no one is denied care based on how much they pay, where they live, or their race, gender, or nationality, would be a huge step forward in reducing the national stain of racial disparities in health.
Bonnie Castillo is the executive director of National Nurses United.